mistralai

Mistral: Mistral Small 3

Mistral Small 3 from MistralAI is designed for text handling with a context length of 32,768 tokens and supports up to 16,384 completion tokens per request. It lacks support for tools or reasoning and does not offer structured output, making it more suited for straightforward text-based tasks. For those seeking a cost-effective solution with a blended benchmark score of 28.8 across multiple independent evaluations, Mistral Small 3 is a viable option. At $0.05 per input token and $0.08 per output token, its pricing makes it accessible for projects requiring substantial text processing while delivering solid performance.

Quality Score
64/100
price + capability + benchmarks
Input Price
$0.05
per 1M tokens
Output Price
$0.08
per 1M tokens
Context Window
32,768
tokens

Benchmark results

Independent, published benchmarks. Blended score 28.8 across 1 benchmark, last refreshed 2026-07-29. How scoring works →

BenchmarkMeasuresScore
EQ-Bench emotional understanding in dialogue 31.8
Model ID
mistralai/mistral-small-24b-instruct-2501
Vendor
mistralai
Released
January 2025
Tokenizer
Mistral
Input Modalities
text
Output Modalities
text
Max Output
16,384 tokens
Tool Calling
not supported
Structured Output
✓ supported
Reasoning Mode
not supported
Vision
text only
Audio
no
Moderated
no

What it costs in practice

Computed from the current $0.05/M input and $0.08/M output rates. Run your own numbers →

JobTokensCost
Summarize a 50-page report 30k in / 1.5k out under $0.01
Classify 1,000 customer emails 500k in / 50k out $0.03
A month of a busy support chatbot 5M in / 2M out $0.41

Similar models

Quick answers

How much does Mistral: Mistral Small 3 cost?
$0.05 per million input tokens and $0.08 per million output tokens.
What is Mistral: Mistral Small 3's context window?
32,768 tokens, roughly 49 pages of text in a single request.
Does Mistral: Mistral Small 3 support tool calling?
It supports structured output.
Can Mistral: Mistral Small 3 process images?
No, it is text-only on the input side.